THESIS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

in

Machine and Vehicle Systems

Structural Safety Design for Real-World Situations Using Computer Aided Engineering for Robust Passenger Car Crashworthiness

L I N U S W Å G S T R Ö M

Department of Applied Mechanics

CHALMERS UNIVERSITY OF TECHNOLOGY

Gothenburg, Sweden, 2013

Structural Safety Design for Real-World Situations Using Computer Aided Engineering for Robust Passenger Car Crashworthiness

LINUS WÅGSTRÖM

ISBN 978-91-7385-904-2

© LINUS WÅGSTRÖM, 2013

Doktorsavhandlingar vid Chalmers tekniska högskola

Ny serie nr 3585

ISSN: 0346-718X

Department of Applied Mechanics

Chalmers University of Technology

SE-412 96 Gothenburg

Sweden

Telephone +46 (0)31 7721000

Cover:

Basic model of how traffic environment affects the vehicle and human

occupants in crashes (left), and simulation setup for oblique crashes described in

Paper III (right).

Chalmers Reproservice

Gothenburg, Sweden 2013

You miss 100% of the shots you don't take.

- Wayne Gretzky

i

Structural Safety Design for Real-World Situations Using Computer Aided Engineering for Robust Passenger Car Crashworthiness

LINUS WÅGSTRÖM

Department of Applied Mechanics

Chalmers University of Technology

Abstract

Road traffic continues to cause more than a million fatalities worldwide every year. Although

many steps have been taken to improve occupant protection in car crashes, challenges still

remain for car designers.

In the present study, real-world data derived from frontal crashes has been used as a base for

identifying crash situations where occupants are severely or fatally injured in cars despite

them having been awarded top-ratings in crashworthiness evaluation tests. One situation

identified is small overlap crashes, where injuries are commonly related to intrusion. Another

is large overlap situations, where injuries are not directly linked to intrusion but rather to

vehicle deceleration and interaction with restraint systems.

The aim of the studies constituting this thesis was to develop design methods for robust

crashworthiness of future passenger cars and propose solutions to mitigate injuries in large

overlap situations. Research was performed using simulation models ranging from simple

mass-spring elements to detailed Finite Element (FE) models of contemporary passenger

cars.

A newly developed methodology has been proposed as a main contribution based on the

research undertaken, in order to provide a comprehensive way of simulating and visualising

structural robustness in car-to-car frontal crashes. The methodology was applied to identify

worst-case scenarios both regarding intrusion (oblique small overlap scenarios) and

deceleration (large, but not full, overlap scenarios). Further development of this methodology

has been proposed in order to address issues of crash compatibility, as well as a tool for

securing robustness in future mass reduction scenarios.

Another contribution is the proposal of an adaptive front structure to reduce passenger

compartment deceleration levels by actively decoupling the front subframe on a

contemporary passenger car in a range of frontal car-to-car crash scenarios. Results suggest a

deceleration reduction potential equivalent to reducing the velocity change in a frontal crash

by up to 44%.

The findings of the present study are compared to previous work and future applications are

suggested.

Keywords: passive safety, crash simulation, structural robustness, frontal crashes, structural

adaptivity, crash compatibility, small overlap crashes

ii

List of appended papers

Paper I

Wågström, L., Thomson, R., Pipkorn, B. (2004) “Structural adaptivity for acceleration level

reduction in passenger car frontal collisions” International Journal of Crashworthiness 9 (2),

121-127.

Paper II

Wågström, L., Thomson, R., Pipkorn, B. (2005) “Structural adaptivity in frontal collisions:

implications on crash pulse characteristics” International Journal of Crashworthiness 10 (4),

371-378.

Paper III

Wågström, L., Kling, A., Norin, H., Fagerlind, H. (2013) “A methodology for improving

structural robustness in frontal car-to-car crash scenarios” International Journal of

Crashworthiness, International Journal of Crashworthiness 18 (4), 385-396.

Paper IV

Wågström, L., Kling, A., Norin, H., Fagerlind, H. (2012) “A Correlation Study for Oblique

Frontal Impacts with Focus on Small Overlap Situations” In: Proceedings of the

International Crashworthiness Conference (ICRASH), Milan, Italy.

Paper V

Wågström, L., Kling, A., Berge, S., Norin, H., Fagerlind, H. (2013) “Adaptive structure

concept for reduced crash pulse severity in frontal collisions” International Journal of

Crashworthiness (published online).

The appended papers were prepared in collaboration with the co-authors. The author of this

thesis was responsible for planning the aim and scope of the papers, carrying out the

numerical simulations, analysing the results, and stating the conclusions.

iii

Table of contents

Abstract ....................................................................................................................................... i

List of appended papers ............................................................................................................. ii

Table of contents ...................................................................................................................... iii

Definitions and abbreviations .................................................................................................... v

Visualisation of vehicle structures ........................................................................................... vii

Preface....................................................................................................................................... ix

Acknowledgements .................................................................................................................... x

1 Introduction ........................................................................................................................ 1

1.1 Basic model for occupant protection in passenger car crashes ................................... 1

1.2 General aim and scope ................................................................................................ 4

1.3 Thesis outline .............................................................................................................. 6

2 Literature review ................................................................................................................ 7

2.1 Current status of passenger car structural design for frontal crashes .......................... 7

2.2 Crashworthiness and robustness.................................................................................. 8

2.3 Crash compatibility ................................................................................................... 10

2.4 Small overlap situations ............................................................................................ 14

2.5 Large overlap situations ............................................................................................ 15

2.6 Adaptive structures .................................................................................................... 16

3 Objectives ........................................................................................................................ 17

4 Paper summaries .............................................................................................................. 18

4.1 Summary of Paper I................................................................................................... 20

4.2 Summary of Paper II ................................................................................................. 21

4.3 Summary of Paper III ................................................................................................ 22

4.4 Summary of Paper IV ................................................................................................ 23

4.5 Summary of Paper V ................................................................................................. 24

5 General discussion ........................................................................................................... 25

5.1 Discussion of results with respect to crash opponent ................................................ 26

5.2 Discussion of results with respect to crash scenario ................................................. 27

5.3 Discussion of results with respect to crash energy .................................................... 28

6 Conclusions ...................................................................................................................... 30

6.1 Lessons learned ......................................................................................................... 30

6.2 Suggestions for future studies and development work.............................................. 32

7 References ........................................................................................................................ 34

iv

v

Definitions and abbreviations

Active safety Measures taken to avoid or mitigate crashes

Aggressivity Measured for a subject vehicle in terms of fatality or injury risk for

occupants in opponent vehicles involved in the same crash

AHOF Average Height of Force

AIS Abbreviated Injury Scale

A-pillar Vehicle structure connecting floor and roof, in front of front doors

B-pillar Vehicle structure connecting floor and roof, rear of front doors

CAE Computer Aided Engineering

CCIS Cooperative Crash Injury Study

Crash Dissipation of vehicle kinetic energy by structural deformation

Crash compatibility Combination of self and partner protection in crashes

Crash pulse Vehicle deceleration time history

ELVA Advanced Electric Vehicle Architectures, research project acronym

Euro-NCAP European New Car Assessment Programme

EVCR Equivalent Velocity Change Reduction

EVERSAFE Everyday Safety for Electric Vehicles, research project acronym

FE Finite Element

FIMCAR Frontal Impact and Compatibility Assessment Research, research

project acronym

Frontal stiffness Relationship between force and displacement during crush of frontal

structures

FWDB Full-width Deformable Barrier

FWRB Full-width Rigid Barrier

GIDAS German In-Depth Accident Study

HIII Hybrid III crash test dummy

HOF Height of Force

IIHS Insurance Institute for Highway Safety

KW400 Work stiffness over the first 400 mm of vehicle front crush

LS-DYNA Finite element solver used for crash simulations

vi

MAIS Maximum Abbreviated Injury Scale

MPDB Moving Progressive Deformable Barrier

NHTSA National Highway Traffic Safety Administration

ODB Offset Deformable Barrier

OLC Occupant Load Criterion

Partner protection A vehicle’s ability to indirectly protect occupants in opponent

vehicles against injuries in car-to-car crashes (i.e., reduction in

aggressivity)

Passive safety Measures taken to reduce consequences of crashes

PDB Progressive Deformable Barrier

Restraint system Interior vehicle system designed to mitigate injuries in crashes, e.g.,

seatbelts, airbags and seats

Self-protection A vehicle’s ability to protect own occupants against injuries in

crashes

SUV Sport Utility Vehicle

Vehicle structure Structural element of vehicle, e.g., body structure or front subframe

Vision Zero Long-term aim for eliminating fatalities and serious injuries in road

traffic

VPI Volvo Pulse Index

WHO World Health Organization

vii

Visualisation of vehicle structures

viii

ix

Preface

The work presented in this thesis was carried out at Chalmers University of Technology in

Gothenburg, Sweden and was divided into two periods.

The first work period was carried out at the Crash Safety Division of the Department of

Machine and Vehicle Systems from 2002 to 2004 under the supervision of Dr. Robert

Thomson and Dr. Bengt Pipkorn and was sponsored by Autoliv Research.

The second work period was carried out at the Division of Vehicle Safety of the Department

of Applied Mechanics from 2011 to 2013 under the supervision of Dr. Hans Norin, Helen

Fagerlind and Anders Kling. All studies during the second period were sponsored by Volvo

Cars and conducted as a part of the Volvo Cars Industrial PhD Programme (VIPP) in

association with SAFER – Vehicle and Traffic Safety Centre at Chalmers, Sweden.

x

Acknowledgements

The creation of this thesis would not have been possible without the support of numerous

colleagues and family members. I would like to mention the people that have played the most

important roles and at the same time apologize to anyone that I may have failed to mention:

My academic supervisor Dr. Hans Norin, for sharing your vast experience within automotive

safety and posing important questions regarding the objective and contents of my work. I

have fully enjoyed the creative atmosphere in our collaboration.

My assistant academic supervisor Helen Fagerlind, for providing a real-world perspective on

my simulation studies as well as caring for the written quality of this thesis and included

papers. You are always a source of optimism and you have fully understood the power of

feedback and encouragement.

My industrial supervisor Anders Kling, and my companion since several years at Volvo

Cars. You are one of the greatest safety enthusiasts around and I have had the privilege of

learning from your knowledge that you share humbly and generously.

My examiner Professor Mats Y Svensson, who I have had the pleasure of interacting with

during both phases of my doctoral studies. Your guidance towards completing this thesis has

been valuable as well as inspiring to me.

My previous academic supervisor Dr. Robert Thomson, for enjoyable collaboration both for

my studies up to the Licentiate degree and later in research projects FIMCAR and

EVERSAFE.

My previous industrial supervisor Dr. Bengt Pipkorn, for guidance towards the Licentiate

degree and sharing your experience within finite element simulations.

Adjunct Professor Dr. Yngve Håland for serving as an inspiration in terms of ingenuity and

authority within the area of vehicle safety.

Professor Per Lövsund, my examiner for the Licentiate degree, for sharing your considerable

experience in vehicle safety research and guidance towards research competence.

My group manager Graeme McInally, section manager Anders Djärv, director Dr. Magnus

Olsson and former director Hans Nyth at Volvo Cars Safety Centre for constant support and

for assisting me in keeping a balance between academic studies and development of even

safer Volvo cars. I would especially like to thank my former group manager Ragnar Crona

for looking beyond daily tasks and supporting me when the second half of my doctorate

studies where set up.

My skilled and hard-working colleagues at the frontal impact analysis group at Volvo Cars

Safety Centre over the years for providing a pleasant and challenging work environment,

especially the people that have been my closest colleagues working with crash simulation:

Peter Appelgren, Anders Carlsson, Petter Hörling, Per-Henrik Myrefelt, Fredrik

Tidborg, Peter Sandqvist, Anders Sandahl, Weijia Wu and Babak Nejabat.

My other colleagues at Volvo Cars Safety Centre, especially section manager Jan Ivarsson,

senior technical leader and adjunct Professor Dr. Lotta Jakobsson, senior technical advisor

Thomas Broberg, technical leader Stefan Berge and technical experts Dr. Johan Jergeus

xi

and Dr. Mikael Ljung-Aust. You all represent extraordinary competence in vehicle safety

engineering that has been a highly valuable asset to me during my studies.

Senior technical leader Dr. Mikael Fermér, serving as a role model in combining academic

merits with industrial excellence. You played an important part in encouraging me to

continue towards this thesis.

All my colleagues at Chalmers University of Technology, especially associate Professor Dr.

Karin Brolin for assistance with Paper III that needed two major revisions before finally

being accepted for publication, Professor Jac Wismans for collaboration in ELVA, and

assistant Professor Dr. Johan Davidsson for assistance regarding teaching.

My former and current doctorate student colleagues, especially Fredrik Jenefeldt,

Dr. Fredrik Törnvall and Dr. Weijia Wu with whom I had the pleasure of sharing office

space with during my first period 2002-2004. The same goes for my colleagues at Injury

Prevention 2011-2013: Dr. Katarina Bohman, Stina Carlsson, Dr. Anna Carlsson,

Kristian Holmqvist, Aleksandra Krusper, Manuel Mendoza-Vazquez, Krystoffer Mroz,

Jóna Marín Ólafsdóttir, Ruth Paas, Isabelle Stockman and Jonas Östh.

The SAFER organisation directed by Anna Nilsson-Ehle, for making my time here pleasant

and inspiring.

My master thesis students Yannick Schroeder, Oscar Centeno, Nilesh Dharwadkar and

Krishna Prashant Adivi that I had the privilege to supervise on their way to becoming

excellent engineers.

Finally, the Wågström clan: my parents Ingrid and Sören; my sister Linda and brother

Pär-Johan with families. Thanks for everything since day one.

My dear family Arezo, Jonathan, Elias and Alicia. This is for you.

Göteborg, September 30, 2013

1

1 Introduction

The World Health Organization (WHO) estimates annual road fatalities to more than 1.2

million globally (World Health Organization 2013), making traffic injuries the leading cause

of death among young people aged between 15 and 29. The WHO reported data from 2006

amounting to more than 100,000 fatalities in India, close to 90,000 in China and more than

40,000 in the United States. Furthermore, the WHO stated in the same report that “while road

traffic death rates in many high-income countries have stabilised or declined in recent

decades, data suggest that in most regions of the world the global epidemic of traffic injuries

is still increasing”.

In the European Union, road transportation claims more than 30,000 lives annually as

reported by the European Commission (2013). Approximately half of the European road

fatalities represent occupants in passenger cars (DaCoTA 2011). The design of passenger cars

must therefore constantly be reviewed to investigate whether there are further steps that can

be taken in order to reduce the number of road fatalities.

A number of countries and organisations have adopted visions for eliminating fatalities and

serious injuries in road traffic, often referred to as Vision Zero (Tingvall and Haworth 1999,

Johansson 2009). It has been identified that countermeasures may be feasible both by

protecting occupants in the event of a crash (passive safety) and by avoiding or mitigating a

crash (active safety) as described by Eugensson et al. (2011).

1.1 Basic model for occupant protection in passenger car crashes

This thesis is focused on passive safety of passenger cars and uses a basic model of how car

occupants interact with the traffic environment via the vehicle in crashes as illustrated in

Figure 1. In the first layer of this model, the outcome for the occupants in a crash is defined

by how they can be protected from crash loads by the vehicle structure and restraint systems.

Restraint systems and vehicle structures are dependent on each other, i.e., without a stable

structural response the restraint systems may not be sufficient to protect occupants against

injuries. Vice versa, the detailed structural response may be less important if occupants are

not restrained to the vehicle in a crash. Therefore, occupant protection is considered to be

based on equivalent shares of structural response and restraint system performance.

In the next layer, the basic model describes vehicle interaction with the crash environment as

illustrated in Figure 1. The crash environment is divided into three categories: crash

opponent, crash scenario and crash energy, all affecting vehicle response in a crash and

defined as follows:

Crash opponent are all types of objects that a car may collide with, e.g., trees, poles,

roadside barriers, animals, trucks, cars or other vehicles.

Crash scenario is a description of how a car interacts with its crash opponent, e.g.,

offset frontal crashes, oblique side crashes, etc.

Crash energy describes, for a situation defined by crash opponent and scenario, the

initial velocity and mass of vehicles and objects involved in the crash.

Any combination of these three crash environment parameters will hereafter be referred to as

a crash situation.

2

Figure 1. Basic model of how traffic environment affects the vehicle and human occupants in crashes.

The objective when designing passenger cars for safety should be that any crash loads

transmitted to the occupant from the traffic environment via the vehicle should be within

human tolerance limits. This objective should ideally be realised for all crash situations, i.e.

combinations of crash opponent, scenario and energy, which the car may be subjected to

during its lifetime. If this cannot be achieved, the most relevant crash situations should be

considered based on the real-world occurrence and corresponding injury risk for each type of

situation. In order to visualise the described traffic environment parameters, they can be

plotted in separate dimensions creating a crashworthiness design volume as illustrated in

Figure 2. This design volume can then be considered to contain all the crash situations that a

passenger car needs to be designed for in terms of structural response, as well as restraint

system performance.

Figure 2. Visualisation of crashworthiness design volume, i.e., three dimensions of traffic environment

parameters: opponent, scenario and energy.

3

The total number of crash situations is inarguably immense and considering every situation

when designing a passenger car is simply not feasible. Therefore, a number of regulatory and

consumer rating load cases have been developed over the years to represent the most relevant

situations. These load cases, or crash setups, are evaluated by means of Computer Aided

Engineering (CAE), i.e., crash simulation, in all phases of product development and by crash

tests later in the development process. If these load cases are selected in an optimal way, all

situations that are relevant in terms of occupant protection for a car’s real-world crash

performance are covered. Feedback on the actual crash performance will be provided during

the car’s life cycle, thus showing strengths and weaknesses in terms of real-world crash

safety performance years after the car model was first introduced on the market. Any design

changes that may be prompted by the real-world crash performance are thus limited to be

implemented late in the product’s life cycle, or to be passed down to the next generation of

vehicles by which time design conditions may have changed.

An alternative approach to awaiting real-world crash data to become available is to use

simulation models to predict real-world crash performance. By using crash simulation, which

is a faster and less expensive method than crash testing, a large number of situations in the

design volume illustrated in Figure 2 can be evaluated. However, this approach relies on

model validity, i.e., if the simulation models cannot be shown to be valid for the complete

evaluation range, results may be inaccurate. Therefore, increased simulation capability for

assessment of real-world crash performance depends largely on improved modelling

techniques such as model refinement and advanced material models including failure

predictions.

Another advantage of using simulation models is that they are well suited for parameter

studies. In this way, individual effects of crash scenario, opponent and energy levels can be

studied. Such relationships may be harder to find in real-world crash data since many

combinations of crash opponent, scenario and energy occur, and confounding factors are

inevitable. Using crash testing to find such relationships is challenging in terms of time and

resources. Furthermore, reconstructing and thoroughly analysing a wide range of real-world

situations by crash testing often requires extraordinary crash laboratory specifications,

including a full range of available crash angles and possibilities to capture film sequences

from underneath vehicles.

4

1.2 General aim and scope

The general aim of this thesis is to contribute to passenger vehicle design methods and

solutions towards eliminating road fatalities and serious injuries as described by Vision Zero

(Tingvall and Haworth 1999, Johansson 2009). This contribution was attempted by using

computer simulation models for use in the design process of passenger cars for improved

real-world crashworthiness. As both structural response and restraint system performance

represent extensive research areas, this thesis is delimited to the design and response of

vehicle structures. Special focus has been directed towards robust structural response, i.e., a

vehicle’s ability to manage a wide range of crash situations without sudden changes of

structural performance.

Furthermore, the studies that comprise this thesis have been delimited to front-to-front car-to-

car crashes at initial oblique angles from -45° to +45° relative to the target vehicle heading

angle in the horizontal plane as illustrated in Figure 3.

Figure 3. Definition of thesis scope with respect to oblique angle in car-to-car crashes.

The delimitation to structural aspects of passive safety is further illustrated in Table 1, where

the scope of this thesis is related to crash scenarios and safety systems.

Part of the aim was to develop methods that are general in their application, i.e., they should

be extendable to a wider scenario range such as side and rear crashes as outlined in Table 1.

Another part of the objective was that methods should be able to support the development of

restraint systems and future decisions on the balance between passive and active safety

systems.

5

Table 1. Overview of crash scenarios and safety systems describing thesis scope limited to methods and

solutions for vehicle structures and frontal crashes.

In order to study and discuss different degrees of horizontal overlap, a definition is needed.

Since passenger cars are designed in various ways, connecting degrees of overlap to a

percentage of vehicle width may not represent structural engagement. Furthermore, when

oblique angles exist, measuring horizontal overlap is subjective and this measurement may

vary during a crash sequence. In this thesis, the extent of engaged structures are therefore

used to define horizontal overlap as illustrated in Figure 4.

Small overlap scenarios are defined as when the frontal structures designed for energy

absorption, such as longitudinal rails (also called side members) or front subframe, are

not engaged.

Moderate overlap scenarios are defined as when the frontal structures designed for

energy absorption on one side of the vehicle are engaged.

Large overlap scenarios are defined as when the frontal structures designed for energy

absorption on both sides of the vehicle are at least partly engaged. If the entire front of

the vehicle is engaged, this is referred to as a full-width scenario.

When analysing real-world crashes as well as crash simulations, it may be difficult to draw

precise borderlines between the different categories as indicated by the gradient shading in

Figure 4.

Figure 4. Definitions of small, moderate and large horizontal overlap in car-to-car frontal crashes.

6

1.3 Thesis outline

This thesis is structured into separate parts as illustrated in Figure 5. The first part serves as

an introduction to the research area and describes the delimitations made in order to move

from the general area of crash safety to the specific area of simulations of frontal car-to-car

crashes. The following part describes a literature review that was conducted in order to

provide an overview of previous work in the area of structural design of passenger vehicles

focused on frontal car-to-car crashes. The literature review also identifies areas of priority

and corresponding research gaps, leading to the objectives of this thesis. In the next part, the

five appended research papers are summarised, focusing on methods and results on the

specific level of simulations of frontal car-to-car crashes (Paper II applies rigid barriers to

estimate energy absorption in frontal crashes). The last part of the thesis discusses the derived

methods and results, and relates them to previous work, as well as possible future

applications and research questions.

Figure 5. Thesis outline.

7

2 Literature review

2.1 Current status of passenger car structural design for frontal crashes

Analysing crash statistics in further detail, it is estimated that approximately half of the fatal

accidents involving car occupants in Sweden occur in frontal crashes (Lindman 2012).

Additional studies show the importance of frontal crashes among situations with severe and

fatal injuries. Over the years 1979 to 2007, NHTSA estimated frontal crashes to account for

44% to 51% of US occupant fatalities (NHTSA 2009). The European FP7 project FIMCAR

reported that frontal crashes represented 57% of UK occupant fatalities 2008-2010 and 32%

of German occupant fatalities 2005-2007 (FIMCAR 2011b).

Regarding frontal crashes, the overall safety level in modern vehicles has been improved

since offset deformable barrier (ODB) tests were introduced. Real-world data show that good

performance in ODB rating tests correlates with reduced injury risk in traffic accidents

(Farmer 2005, Kullgren et al. 2010). However, there are still situations where improvements

can be made which will be described in the following sections.

A specific subset of frontal crashes is frontal car-to-car crashes. In these situations, the crash

performance of frontal structures is tested in real-world situations versus other designs.

European data from frontal car-to-car crashes suggests that the share of occupant fatalities

occurring in frontal car-to-car crashes is 12% in Germany and 23% in the UK (FIMCAR

2011b), illustrated in Figure 6. By assuming that approximately 15% of car occupant road

fatalities occur in frontal car-to-car crashes, this type of situation is estimated to account for

more than 2,000 European road fatalities annually.

Figure 6. Approximate distribution of EU road fatalities, data from FIMCAR (2011b).

Severe injuries to occupants in frontal crashes, as defined by a Maximum Abbreviated Injury

Scale (MAIS) score of three or higher (Sherwood 2009) have been shown to be strongly

linked to passenger compartment intrusion. Furthermore, similar correlation with intrusion

has been found for Injury Severity Scores (ISS) above 25 (Conroy et al. 2008). These

findings support the need for improved car structural integrity, i.e., minimising intrusion into

the passenger compartment. When structural integrity is compromised, injury risk for car

occupants increase both from direct contact with intruding structures, as well as by

influencing the effect of available restraint systems (Brumbelow and Zuby 2009). These

findings from real-world situations propose structural integrity as a crucial performance

measurement for passive safety of passenger cars.

8

2.2 Crashworthiness and robustness

In order to ensure that the vehicle structural response will support the occupant restraint

systems in all relevant situations in the design volume outlined in Figure 2, some

measurement of structural response is needed. Whether it is deceleration, intrusion or another

measure, this measurement should be possible to record and visualise over the entire design

volume in order to compare design options. Therefore, definitions of optimised design

compared to robust design used in this thesis are required.

Consider a system where one output variable describes the system performance as a function

of one input variable as illustrated in Figure 7. The objective of the system is to minimise the

output variable, e.g., passenger compartment intrusion or deceleration which must be

achieved for all input variable settings, e.g., horizontal overlap, within its lower and upper

bounds as illustrated in Figure 7. Furthermore, all input settings must yield output below the

maximum allowable response indicated by the horizontal line in Figure 7

In this example, the definition of optimal input is the input leading to minimum output.

However, if this optima is associated with a small range of input variable settings, a large

output variation may be the result of a minor input variation as described in further detail by

Lee and Park (2001). Therefore, a more desirable input may be one resulting in a higher

minimum performance but lower performance variation.

The definition of a robust solution is a solution where large variations in input result in small

variations in output, as illustrated in Figure 7. If the input is considered an adjustable

variable, the input should be controlled towards an optimal or robust solution. If the input

variable is to be considered a random variable to be handled by the system, the design should

be made in such a way that no input variable settings result in unacceptable performance as

illustrated in Figure 7. In order to address the problem of robustness and unacceptable

performance, a proposal for a modified design is given by the dashed curve in Figure 7. On

the one hand, this design provides a higher minimum performance. On the other hand, the

modified design provides a more robust system, i.e., smaller variations in the output variable

when the input variable is changing. In addition, the modified design does not violate the

maximum allowable response limit as indicated in Figure 7.

Figure 7. Schematic figure showing principles of optimisation and robustness.

9

As described in previously published work on this subject (Marklund and Nilsson 2001,

Craig et al. 2005, Lönn et al. 2009, Lönn et al. 2010, Lönn et al. 2011), studying both

structural optimisation, as well as robustness requires well-defined measurements of

performance. The performance is often called response values (one output variable) or

response surfaces (two or more output variables). In crashworthiness, injury mechanisms are

related to passenger compartment intrusion, intrusion velocity and/or deceleration (Conroy et

al. 2008, Sherwood 2009). Therefore optimisation or robustness response surfaces could

possibly be based on these measurements. However, since car-to-car frontal crashes can be

described as highly non-linear systems, optimisation was not attempted within in the scope of

this thesis. The input variables in car-to-car crashes may be regarded as random to some

extent; an appropriate objective for structural design could be the robustness of the intrusion

and deceleration response of vehicles involved. This means, designing vehicle structures that

exhibit small variations in structural response although variations in input as described by

parameters such as oblique angle and horizontal overlap may exist.

10

2.3 Crash compatibility

Aiming at a robust structural response, an important aspect of occupant protection in

passenger car frontal car-to-car crashes is the opponent encountered. This topic, called crash

compatibility, has been thoroughly studied in various research initiatives over the years. The

main findings of this work are presented in the present section, giving an outline of the

complexity of the subject.

Crash compatibility is considered a combination of both self and partner protection, i.e.,

protecting occupants in the own vehicle as well as occupants in opponent vehicles (FIMCAR

2011a). This definition is needed since protection of occupants in one vehicle should not be

achieved by reducing occupant protection in opponent vehicles.

Compatibility was considered a prioritised research area in the late 1990s and early 2000s. It

was shown that light truck vehicles were over-represented in US car-to-car opponent fatality

statistics and the concept of vehicle aggressivity or partner protection was commonly used to

describe the problem (Gabler and Hollowell 2000, Austin 2005, Huang et al. 2011). This

means that some vehicles cause a disproportionately large number of severe injuries or

fatalities in opponent vehicles in car-to-car frontal crashes. Crash incompatibility and

aggressivity issues were observed also in UK data (Edwards et al. 2001), French data

(Delannoy and Diboine 2001), Canadian data (Fredette et al. 2008) and Japanese data

(Mizuno and Kajzer 1999).

Moreover, the compatibility issue was also found in Swedish accident statistics and attempts

were made to separate the effects of mass and structure by observing police reported two-car

crashes (Kullgren et al. 2001). Kullgren et al. found SUVs to be considerably more

aggressive than the average car. Multibody simulation models were developed as an approach

to compatibility research for estimating effects of fleet changes. (Buzeman-Jewkes et al.

1999, Buzeman-Jewkes et al. 2000). Similar approaches were later attempted that underlined

the usability of such simplified vehicles in order to perform a large number of simulations to

provide a basic understanding of the complex system of a large car fleet (Jenefeldt and

Thomson 2004, van der Zweep et al. 2005, Watanabe et al. 2005). Other researchers used FE

models and crash testing to develop front structure concepts with multiple load paths in order

to address compatibility issues (Fujii et al. 2003, Saito et al. 2003). Attempts were also made

to make car consumers increasingly aware of the safety issues related to vehicle crashes

between vehicles of different size and weight (The Insurance Institute for Highway Safety

2005, The Insurance Institute for Highway Safety 2009)

Starting in 2003, a European consortium consisting of safety research institutes, authorities

and vehicle manufacturers joined forces in the project VC-Compat to find solutions to crash

compatibility issues. In the final VC-Compat technical report (VC-Compat 2007), it was

estimated that approximately 1,000 lives could be saved annually on European roads by

improved crash compatibility. It was identified that structural interaction and compartment

strength are key issues to be addressed in order to improve crash compatibility. However, no

consensus could be reached on standardised test procedures in order to introduce legislative

or consumer rating tests for enhanced crash compatibility.

In order to advance further towards better crash compatibility in frontal car-to-car crashes,

VC-Compat was followed by the 2010-2012 research programme, Frontal Impact and

Compatibility Research (FIMCAR). The starting point of FIMCAR was that crash

compatibility consists of both self and partner protection and the consortium finally reached a

11

recommendation to introduce a full-width test procedure for Europe based on priority of

structural interaction and restraint system performance (FIMCAR 2011e).

The importance of structural interaction appears to be a common conclusion in many of the

studies conducted on crash compatibility. In the United States, a voluntary agreement was

initiated by the Alliance of Automobile Manufacturers in 2003 (Barbat 2005). This

agreement was based on placing frontal energy absorbing structures in a common interaction

zone at 16 to 20 inches above ground level. Later studies from NHTSA (Greenwell 2012) and

IIHS (Baker et al. 2008, Teoh and Nolan 2012) all indicate improvements to crash

compatibility after the agreement on alignment of vehicle front structures had come into

effect.

The results so far from the voluntary agreement suggest that a relatively simple target of a

common interaction zone has been effective in addressing crash incompatibility. However,

developing test procedures using crash barriers to objectively measure appropriate structural

interaction frontal force levels has proved to be difficult.

One approach to measuring structural interaction properties of a vehicle front structure is to

measure the height of force (HOF) time history from a full-width barrier. Alternatively, the

average height of force (AHOF) has been proposed as an average over a certain evaluation

period in time. The applicability of HOF and AHOF were investigated thoroughly by

researchers from the United States (Verma et al. 2004, Subramaniam et al. 2007, Nusholtz et

al. 2009, Brewer et al. 2011) as well as Japan (Mizuno et al. 2005, Watanabe et al. 2005,

Hirayama et al. 2007, Uwai et al. 2007, Yonezawa et al. 2008). Some criticism was raised

regarding the reproducibility of AHOF and possible velocity dependence. Regardless of

evaluation method, the height at which forces are transmitted appears to be of great

importance to car-to-car crash compatibility. Vertical misalignment has been shown to be

unbeneficial for energy absorption and deformation modes which may increase injury risk to

occupants in both vehicles involved in frontal car-to-car crashes (Baker et al. 2008, Mizuno

and Arai 2010).

A second approach suggested is to use the deformation pattern of offset deformable barriers

to assess structural interaction. Aiming at increasing the test severity for small vehicles and

avoiding bottoming-out of the deformable barrier honeycomb block, the Progressive

Deformable Barrier (PDB) was developed for offset crash tests (Delannoy and Diboine 2001,

Delannoy et al. 2005, Delannoy et al. 2007). Additional studies were conducted regarding the

feasibility of attaching the PDB to a trolley, creating a so-called Moving Deformable Barrier

or MPDB (Schram and Versmissen 2007, Versmissen et al. 2007). The MPDB has the

potential to further increase the crash severity for light cars which leads to discussions

regarding the initial velocity and trolley mass setup for this specific load case. If a MPDB test

is always run with a fixed trolley mass and initial velocities as proposed by FIMCAR

(FIMCAR 2011d), this test method could potentially decrease the self-protection of heavier

vehicles since the crash severity will be reduced for this type of cars. The PDB was further

studied by NHTSA (Meyerson et al. 2009) where the PDB was once more proposed as a tool

for assessing partner protection. The PDB has been proposed to be introduced into regulation

as an update to the UNECE R94 regulation (Chauvel et al. 2011). The crash compatibility

assessment potential of the PDB was recognised by FIMCAR, but could not be recommended

for regulatory testing since barrier evaluation metrics were identified to require further

development and validation (FIMCAR 2011c).

12

A third approach is a full-width deformable barrier (FWDB) that could potentially replace a

full-width rigid barrier (FWRB) in tests for high deceleration response (Edwards et al. 2003a,

Edwards et al. 2003b, Edwards et al. 2007, Edwards 2009). By assessing the barrier face

deformation, structural interaction may also be assessed. Additional studies were performed

by Arai et al. (2007) which supported the FWDB as a tool for assessing structural interaction.

Similarly to the PDB, FIMCAR found that although the FWDB could potentially be used for

assessment of structural interaction, these assessment metrics needed further development

(FIMCAR 2011e). FIMCAR however recommended the FWDB to replace the FWRB in

regulatory testing mainly based on indications that the FWDB yields an occupant

compartment deceleration response more representative of real-world situations in the initial

stage of the crash compared to the FWRB (FIMCAR 2011e).

Once structural interaction is improved, the next level of improved crash compatibility is

believed to be stiffness matching, i.e., harmonising frontal force levels for dissimilar vehicle

mass categories. One approach to this could be to base frontal force levels on average vehicle

mass rather than the actual vehicle mass. Based on FE simulations of frontal car-to-car

crashes, it has been suggested that matching the stiffness of a lighter vehicle up to the level of

a 30% heavier car may only have minor effects on the crash pulse shape (Volvo Car

Corporation et al. 2010). It has been shown however, that when stiffness of a lighter vehicle

is matched to a heavier crash opponent, dummy injury values in car-to-barrier tests are

increased significantly (Watanabe et al. 2005).

Matching the frontal stiffness of a heavier vehicle down to the level of a lighter vehicle

requires extended front structure deformation length to maintain energy absorption for self-

protection in the heavier vehicle. Without structure extension, such a design change could

reduce intrusion in the opponent car below a certain impact velocity, but increase intrusion at

higher impact velocities (Hirayama et al. 2007). Subramaniam et al. (2007) explored the

effect of modifying the initial stiffness of a light truck vehicle to match that of a car crash

opponent, and increasing the stiffness in the later phases of crush to compensate for the lost

energy absorption. The result of this modification was a deteriorated car-to-barrier

performance in terms of intrusion as well as deceleration and dummy response. It was

recommended that a wide range of crashes should be evaluated for effects on occupant

protection before any stiffness matching regulations are implemented.

A concrete proposal on how to implement stiffness matching of frontal structures in

passenger vehicles was presented by NHTSA and called KW400 (Patel et al. 2007). This

metric was established as an attempt to measure the frontal stiffness during the first 400 mm

of deformation in a FWRB crash, and was used for stiffness matching studies (Hirayama et

al. 2007, Subramaniam et al. 2007, Nusholtz et al. 2009). One important issue that was raised

was the sensitivity of KW400 to the starting time of the signals (also known as time zero).

Since KW400 is calculated from the barrier force starting at 25 mm of vehicle displacement,

the initial vehicle-to-barrier contact becomes an important parameter, rewarding a low initial

barrier force response.

It has been noted that improved passenger compartment integrity does not necessarily mean

higher stiffness of the front structure (Lund and Nolan 2003). Overall, the studies performed

on stiffness matching underline the balance between self-protection and partner protection.

Given the available front deformation distance, it appears inevitable to match frontal force

levels of frontal structures without affecting this balance.

13

Adding to the complexity of the subject, real-world data suggest that incompatibility in

frontal crashes can exist even if vehicles are identical. In FIMCAR (2011a), an example of

this from the Great Britain Cooperative Crash Injury Study (CCIS) was given. In this case,

two vehicles of the same make and model were involved in a head-on crash at approximately

50% overlap. As illustrated in Table 2, the two cars exhibited significantly different structural

responses, with up to seven times greater dashboard intrusion in one of the cars which was

also reflected in the injuries sustained by the drivers.

Table 2. Overview of the outcome in a head-on crash involving two cars of same model.

Adapted from FIMCAR (2011a).

Model year 2002 2001

Kerb mass 1,423 kg 1,384 kg

Overlap 51% 50%

Equivalent test speed

based on deformation 26 km/h 46 km/h

Dashboard intrusion 190 mm 900 mm

Footwell intrusion 170 mm 1,180 mm

Driver injury level MAIS 2 MAIS 5

The conclusion from studies of previous work within the research area of crash compatibility

between passenger vehicles in frontal crashes is that there is a range of factors that affect

structural response and thereby occupant injury risk. Looking solely at real-world crash data

appears to be incomprehensive for understanding the mechanisms that affect structural

integrity and robustness. Crash simulation could be one tool to isolate unbalance from crash

scenario, as attempted using modified public domain models (Thomson et al. 2008).

In line with this, detailed efforts to describe crash incompatibility in the case of dissimilar

vehicles will not be attempted in this thesis. Instead the incompatibility that arises from the

crash scenario will be explored. In order to protect occupants in frontal crashes, vehicles

should be developed to provide a robust and predictable structural response in frontal crashes

independent of crash opponent.

14

2.4 Small overlap situations

Another key issue for the design of vehicle structures for real-world crashworthiness in

frontal crashes are situations where structures intended for energy absorption are not

engaged. One such situation occurs when the front of a vehicle is loaded outboard of the front

structure, often called severe frontal collision with partial overlap (Planath et al. 1993) or

simply small overlap crash (Lindquist et al. 2004), as illustrated in Figure 4.

The real-world significance of small overlap crashes is not a new phenomenon. The

importance of designing vehicles for this load case was described by Planath et al. (1993)

where a test method with 20 to 40% horizontal overlap against a fixed rigid barrier at initial

velocities up to 65 km/h was proposed. Accident data from the early 1990s suggested that the

percentage of moderately and severely injured drivers was higher in crashes with an overlap

below 30% than in crashes with an overlap of more than 30%. (Kullgren and Ydenius 1998).

Further indications of the importance of small overlap crashes was presented by Lindquist et

al. (2004), where small overlap crashes accounted for 48% of the fatalities to belted

occupants in a data set of frontal crashes in Sweden.

Small overlap situations were given renewed focus when the IIHS highlighted this type of

situation in the early 2010s. When observing real-world frontal crashes involving vehicles

awarded with good ratings in the IIHS test programme for frontal crash protection

(Brumbelow and Zuby 2009), small overlap crashes was one type of scenario where

occupants sustained severe injuries. The strong link between occupant compartment intrusion

and injury severity in small overlap crashes was presented in a later study by the IIHS

(Sherwood 2009). This lead to the development of an IIHS small overlap rigid barrier load

case, described in detail by Sherwood et al. (2013). The IIHS also demonstrated that the test

method shows acceptable repeatability, indicating that vehicles with poor structural

performance in this load case show the largest test-to-test variations in terms of intrusion

(Mueller et al. 2013).

NHTSA also suggested that small overlap crashes should be a priority area for future research

(Rudd et al. 2009), followed by an additional study where it was demonstrated that small

overlap crashes frequently produce oblique kinematics, and the interaction along the side of

the struck vehicle increases the risk for injuries from outboard components such as the door

and A-pillar (Rudd et al. 2011).

Several studies have thus indicated that these situations are critical for reducing severe and

fatal injuries. There are, however, studies suggesting that “the small overlap is at worst a

moderately dangerous crash in the overall scheme of frontal crashes” (Scullion et al. 2010).

In a later study (Kühn et al. 2013), the German Insurers Accident Research supported the

IIHS finding that approximately 25% of frontal crashes can be characterised as small overlap

situations. In this German dataset, small overlap crashes were found to represent a small

number of fatalities but a large number of serious (AIS2+) injuries to the lower extremities.

Although the representativeness of the IIHS small overlap barrier crash test appears to be a

subject for debate, real-world data suggest that small overlap situations should not be

neglected when striving for a vision of zero fatalities and serious injuries. Further support for

a small overlap barrier load case to represent small overlap car-to-car scenarios was given by

Jakobsson et al. (2013a) combined with an overview of vehicle design changes for small

overlap situations (Jakobsson et al. 2013b).

15

2.5 Large overlap situations

In addition to crash compatibility issues and small overlap crashes, there appears to be issues

related to the crash pulse or restraint system in frontal crashes with modern cars. In a US

study on the types of frontal crashes that cause serious injuries and fatalities to belted front-

seat occupants in passenger cars, a considerable portion of serious injuries occur in frontal

crashes despite good structural integrity (Brumbelow and Zuby 2009). In that study of cars

that had been awarded good ratings in the IIHS frontal moderate (40%) overlap test, it was

shown that many belted occupants still sustain severe injuries in frontal crashes without

significant vehicle passenger compartment intrusion. This means that even without intrusion,

occupants may be injured from contact with restraint systems or car interior, i.e., injury

mechanisms are related to the crash pulse rather than intrusion.

Similar findings were presented in the European research project FIMCAR (2011a), where

datasets of frontal crashes involving R94-compliant vehicles from Great Britain (Cooperative

Crash Injury Study, CCIS) and Germany (German In-Depth Accident Study, GIDAS) were

analysed. The study showed that approximately 40% of MAIS 2+ injuries and 30% of fatal

injuries suffered by occupants occurred in crashes with more than 75% frontal overlap

(compare to Figure 4), and it was suggested that compartment intrusion may not be the direct

cause of injury. Besides improving the functionality and robustness of restraint systems,

injuries related to the crash pulse can potentially be addressed by the vehicle front structural

response.

Current vehicle structures may exhibit stiffer response in frontal crashes compared to older

vehicles, especially in the late phases of the crash pulse as a direct effect of improvements in

terms of intrusion (Nolan and Lund 2001, Samaha et al. 2010). It is therefore important to

balance high-deceleration load cases with offset load cases with significantly different crash

pulse shapes and structural loading (FIMCAR 2011a).

16

2.6 Adaptive structures

Adaptive structures could be one way to optimise the structural response in the wide range of

crash scenarios that passenger vehicles encounter. In frontal crashes, it has been suggested

that adaptive structures can be used in order to affect the deceleration response. Witteman

and Kriens (2001) followed by Witteman (2005) suggested “high-low-high” deceleration

pulses to be optimal based on occupant response simulations of crashes with velocity change

of 56 km/h or greater. These deceleration pulses were proposed to be accomplished by

friction forces applied to steel cables that had the additional benefit of being able to transfer

loads from the struck side of a vehicle to the non-struck side.

An alternative approach to achieving “high-low-high” deceleration pulses in a passenger car

has been proposed (Motozawa and Kamei 2000, Motozawa et al. 2003), where axial buckling

is followed by bending of the main energy absorbing members. A practical solution for the

required operational volume for such a system was, however, not presented.

Pipkorn et al. (2005) recommended implementing variable crush force in passenger cars by

pressurising vehicle longitudinal frontal members. Since the additional volume required for

the pressurised frontal members to function may be minor, this solution may be more

efficient in terms of packaging space than the concept proposed by Motozawa et al. (2003).

Any solution as to how the Pipkorn et al. proposal would be applied to production vehicles

was not presented, although the mass-reducing potential based on increased force levels from

pressure in such members was highlighted. In a subsequent study, Pipkorn and Kullgren

(2009) again showed that pressurising thin-walled tubular structures can significantly

increase the crush force and energy absorption and considerably reduce occupant injury risk

as measured by HIII dummy readings. Furthermore, a comprehensive review on adaptive

vehicle structures was made by TRL (Thompson et al. 2007), where altering the frontal force

level was identified as one of the key principles for adaptive structures.

Other applications of adaptive structures were proposed by Pipkorn et al. (2007) such as

attaching an external inflatable airbag to the front structure of an SUV. This was

demonstrated to increase structural interaction by creating a lower load path into the sill of a

passenger car in side crashes where the SUV is the bullet vehicle. Two types of adaptive front

structures, fixed and extendable, were investigated using simplified, two-dimensional,

simulation models by Elmarakbi and Zu (2006). The study predicted that improvements, in

terms of injury risk, related to both intrusion and deceleration could be accomplished by

adaptive front structures.

Pyrotechnically controlling the stiffness of load-carrying front structures of passenger cars in

front-to-side crashes were investigated by Ostrowski (2007). In that study, an adaptive

structure was shown to decrease the front-end crash stiffness of the bullet vehicle and to

extend the crushing distance when needed. Another thorough review on adaptive structures

for improved crashworthiness was made by Khattab (2011), where extendable add-on energy

absorbers were also suggested. Additionally, Pipkorn et al. (2011) suggested further

applications of adaptive structures in passenger cars. In that study, the balance between

forward vision and A-pillar load capacity was suggested to be improved by introducing

expandable A-pillars.

17

3 Objectives

The general aim of this thesis is, as previously described, to investigate novel methods based

on computer simulation models that can be used in the design process of passenger cars for

robust real-world crashworthiness. Supporting Vision Zero through vehicle design requires a

holistic view on safety, beyond the load cases defined by legal requirements and consumer

rating programmes. Given the level of reliability that crash simulation models have reached

through constant development over several decades, these models today represent valuable

tools for understanding and proposing countermeasures for the structural challenges seen in

real-world crash situations.

One specific objective of this thesis was therefore to develop tools for engineering of safe

future passenger vehicles. These tools should be applicable for addressing crash

compatibility, small overlap situations and large overlap situations with focus on robust

structural response. A second specific objective of this thesis was that the developed tools

should be used to explore the applicability of adaptive front structures for crash severity

reduction in frontal car-to-car crashes.

18

4 Paper summaries

The individual papers are linked by their connection to real-world data as illustrated in Figure

8. All five appended papers are summarised in the following section and their relative

coherence is explained here.

Two major crashworthiness issues were found in literature regarding real-world frontal car-

to-car scenarios where injuries occur in modern vehicles. The first issue concerns small

overlap situations, where intrusion appears to be the main cause of injuries. The second

concerns large overlap scenarios where intrusion is not necessarily the major cause of injuries

but rather the deceleration of the passenger compartment.

Based on these two problems seen in real-world crash data, the work was divided into two

paths – one for large overlap situations and one for small overlap situations. Paper I served as

an initial study of structural adaptivity in frontal crashes and Paper II followed up on those

findings by exploring effects on the crash pulse with a public domain FE model. Paper III

laid a foundation for a methodology to be used for structural robustness in frontal crashes

regardless of scenario. Paper IV used the methodology from Paper III to examine and classify

small overlap situations by validating the FE model against full-scale crash tests. To finally

bring both halves of the thesis together, Paper V used parts of the methodology from Paper

III to set up a large number of crash simulations to explore a specific concept for structural

adaptivity, a detachable front subframe.

Figure 8. Overview of study coherency with appended papers in roman numerals.

The studies conducted were divided according to the respective issue seen in real-world crash

data and their nature, i.e., being a method or a proposed solution. For small overlap scenarios,

any specific solutions to address the issues seen in real-world crashes were not proposed.

However, for the issues with larger overlap, i.e., deceleration-related injuries, adaptive frontal

structures were proposed as a possible solution that could potentially address this type of

situation.

19

All five papers can be related to the crashworthiness design volume in Figure 2. This is

illustrated in Figure 9 where each point represents a crash situation evaluated by means of

simulation. Figure 9 shows that only Paper I attempts to assess the effect of the crash

opponent, whereas Papers II to V are focused on crash scenario or crash energy. Paper II

applies rigid barrier simulations for estimation of energy absorption relevant to car-to-car

crashes with satisfactory structural interaction.

Figure 9. Overview of which crash environment parameters were addressed in the appended papers.

Each point represents an evaluated crash situation.

A central concept in Papers III to V is crash pulse severity metrics. These are based on

acceleration signals of the sill structures close to the B-pillars. These signals were used to

estimate the crash pulse severity based on two simplified models, the Volvo Pulse Index, VPI

(ISO 2012) and the Occupant Load Criterion, OLC (Stein et al. 2011). Both models are

attempts of generically measuring the restraint forces that the driver is subjected to during a

crash, based on deceleration only. Each model uses the occupant displacement relative to the

vehicle in the longitudinal direction. Furthermore, both models assume an initial phase of

free-flying motion without any occupant deceleration at relative displacement less than 30

mm for VPI and 65 mm for OLC, as illustrated in Figure 10. However, the model responses

following the initial slack represent fundamentally different assumptions. VPI assumes a

linearly increasing occupant deceleration of 0.25 g/mm of relative displacement, without any

limitation on occupant deceleration or relative displacement. The OLC instead assumes a

maximum relative displacement of 300 mm and a constant occupant deceleration up to this

point. This means that the OLC model assumes a perfectly adaptive restraint system that will

always utilise the available interior distance. The VPI model, on the other hand, simulates a

non-adaptive restraint system based on chest decelerations measured in crash test dummies in

physical tests.

Figure 10. Characteristics of the crash severity indicators VPI and OLC.

Occupant deceleration plotted vs. occupant displacement relative to vehicle.

20

4.1 Summary of Paper I

4.1.1 Introduction

In-depth studies of crash pulses from real-world frontal crashes have shown a correlation

between acceleration levels and injury risk, indicating that high vehicle acceleration during

frontal crashes increases the risk of long-term consequences for the occupants. The only way

for a vehicle manufacturer to affect the crash pulse is to change the characteristics of the

energy absorbing parts of the car. Consumer tests such as Euro-NCAP have prompted car

manufacturers to design vehicles with high intrusion resistance. However, in low severity

crashes, these structures could potentially subject car occupants to higher accelerations than

what would be possible if the stiffness could be adapted to crash severity. Suggestions to

design adaptable frontal structures have been found in the literature. This study has sought to

find guidelines of how to select deformation characteristics that lead to less harmful

acceleration pulses during low-speed crashes while maintaining intrusion resistance at higher

crash velocities in frontal crashes.

4.1.2 Method

Car-to-car crashes were simulated using mass-spring models assuming linear frontal stiffness.

Typical values for vehicle frontal stiffness were chosen after separating the car fleet into two

categories, where one category consisted of cars with a test weight below 2000 kg and the

other category consisted of cars with a test weight above 2000 kg. Crash test data was used to

approximate frontal stiffness.

To simulate full-frontal crashes between cars of dissimilar mass or frontal stiffness, three

vehicle classes were defined. Using all combinations of car classes and three closing

velocities ranging from 40 to 120 km/h resulted in a total of eighteen simulations that were

run with the baseline configuration, i.e., constant stiffness. The results of these simulations in

terms of maximum deceleration were then compared to the results using an adaptive

deformation system, where the frontal stiffness was minimised for each level of kinetic

energy.

4.1.3 Results

All peak acceleration values were decreased as a result of the adaptive system. Peak

deceleration values were on average reduced by 14% at the highest closing velocity of 120

km/h. When the closing velocity was decreased to 80 km/h, the corresponding reduction was

43% and at 40 km/h, the average peak value was reduced by as much as 73%. It was

observed that at the lowest closing velocity, the calculated adaptive stiffness was less than

10% of the original value.

4.1.4 Discussion

Systems including an adaptive frontal stiffness could require pre-crash sensors that can

provide information both regarding the state of the own vehicle, as well as the collision

object. Structural adaptivity could be accomplished by decreasing or increasing the internal

force levels of the deforming vehicle parts. This should be executed in a manner that vehicle

occupants are guaranteed an acceptable level of protection in case of undesired system

response. It was suggested that a fully-developed adaptive system would have the potential to

decrease deceleration levels in low severity accidents, as well as increasing the deformation

energy absorption in high velocity crashes.

21

4.2 Summary of Paper II

4.2.1 Introduction

In passenger-car crashes, frontal crashes are the most frequent accident type. Modern cars are

able to maintain structural integrity even in high velocity crashes and thereby reducing the

risk of severe and fatal injuries. Indications of safer cars were found in crash statistics in

terms of lower injury risk to all body regions, with one alarming exception: injuries to the

neck. Studies of crash pulse recorder data have suggested mean deceleration as a candidate

crash-severity measure for AIS1 neck injuries.

It has been suggested that greater passenger compartment integrity may have made vehicles

less forgiving in terms of deceleration-related injuries such as long-term neck injuries. To

reduce harmful crash pulses caused by stiff frontal structures, it is suggested that the optimal

frontal structure stiffness should be adapted to the crash situation. The logical approach to

achieve lower deceleration is to weaken the engaged structural members. By altering the

structural deformation characteristics in suitable locations of a passenger car front, the

response could be made sufficiently supple for a low-velocity crash while the front-end

stiffness is maintained or increased in a high-velocity crash.

4.2.2 Method

A public domain full-vehicle LS-DYNA FE model of the passenger car model Geo Metro

was used to simulate full-width crash and offset crash with a rigid barrier that covered 40% of

the maximum vehicle width. In four preliminary simulations, the longitudinal rails were

identified to represent the largest portion of absorbed internal energy for impact velocities of

both 32 km/h and 56 km/h.

To study the effect on crash pulse characteristics and the possible implications of structures

with alternating structural strength, the material of the longitudinal rails was given three

levels in subsequent simulations at a range of initial velocities chosen from 16, 32, 48 and 64

km/h. The passenger compartment was modelled in two variants, a rigid compartment and the

original deformable compartment.

4.2.3 Results

Minimising passenger compartment intrusion, assessed by a rigid compartment model, will

potentially reduce the crash pulse duration and thus increase mean deceleration. The effect of

front longitudinal members crush strength on the crash pulse depends largely on the

geometric constraints in the engine compartment and the kinetic energy to be absorbed in a

crash. Reducing the yield stress of longitudinal members can only reduce the peak

deceleration if this peak is associated with structural deformation and not engine contact with

a stiff passenger compartment.

4.2.4 Discussion

To design adaptive frontal structures that have a significant effect on the crash pulse,

strategies for affecting the global load paths must be investigated. This study suggests that

changing the strength of the most significant energy absorbing structural members would

only affect the crash pulse to a limited extent.

22

4.3 Summary of Paper III

4.3.1 Introduction

From data collected in real-world frontal crashes involving new vehicles, it has been

suggested that consumer rating crash tests have encouraged improvements to passenger

vehicle structural crashworthiness. However, there still seems to be room for further

improvement in so-called small overlap conditions, i.e., where vehicles are involved in

frontal crashes without engaging the main frontal crash absorbing structures.

It has also been shown that the available coding standards for crush damage in passenger

vehicle crashes do not always capture differences in crash configuration, making detailed

parameter analyses of car-to-car crash configurations a difficult task. Additional approaches

such as computer simulation are therefore needed for understanding how different crash

scenarios are linked to vehicle structural performance and possibly occupant injury risk.

Based on these findings from previous studies and real-world crash data, it is suggested that

improved crash simulation techniques should be developed in order to better understand the

structural mechanisms that lead to large passenger compartment intrusion or deceleration in

frontal car-to-car crash scenarios. The aim of this particular study was therefore to develop a

methodology for identifying dimensioning frontal car-to-car crash scenarios by assessing

crash configuration parameters that influence structural response.

4.3.2 Method

A full-vehicle FE model was validated in terms of intrusion and deceleration response in

frontal crashes and used to establish a car-to-car crash simulation model with two identical

passenger cars. The car-to-car FE model was employed for a parameter study including 378

simulations on how the crash setup affects passenger compartment intrusion and deceleration.

Based on this output, a set of scenarios with outstanding properties in terms of structural

response were defined and considered candidates for crash scenarios that should be used for

dimensioning of car structures.

Since the FE model used for the parameter study cannot be validated by physical testing in all

of the scenarios defined by the simulation matrix, a few of the most noteworthy crash setups

have been selected for further validation work as they fell outside the scope of the study. By

improving numerical robustness and validity of FE models, the methodology would be

suitable for comprising part of a comprehensive toolbox for ensuring robust response of

vehicle structures.

4.3.3 Results

The intrusion area that displayed the best correlation with other intrusion areas was the

central A-pillar intrusion which was therefore suggested to represent the overall intrusion

levels in each car. At 15° oblique angle and 1,200 mm lateral offset scenarios, this intrusion

was more than three times greater in one of the identical cars. The greatest crash pulse

severity was found at scenarios around 300 mm lateral offset and 0 or 5° oblique angle.

4.3.4 Discussion

The purpose of the methodology presented in this study was to establish a tool for structural

robustness in the development process of passenger vehicles. The proposed compatibility

domain lends itself to visualising structural differences in car-to-car crashes regardless of the

vehicles involved. In an extended application, this methodology could be used to compare the

relative importance of different aspects of incompatibility, e.g., studying when a structural

advantage is cancelled by an unbeneficial car-to-car crash scenario.

23

4.4 Summary of Paper IV

4.4.1 Introduction

It is estimated that approximately half of the European Union road transportation fatalities are

occupants in passenger cars, and that about half of these occur in head-on crashes. Consumer

rating programmes using ODBs performed in both Europe and the United States have

encouraged car designs with improved passenger compartment integrity in order to increase

self-protection. However, fatalities and severe injuries still occur and it is therefore important

to study these types of crashes. Data from Europe as well as the United States suggest

situations were defined as situations where front structures designed for energy absorption are

not engaged as major load paths, often called small overlap frontal crashes.

In addition to designing vehicles for self-protection, consideration should also be made for

partner protection. Incompatibility in two-vehicle frontal crashes is characterised by

differences in injury risk to occupants in one vehicle compared to the other vehicle. These

differences can be caused by both occupant and vehicle dissimilarities, but are normally

discussed in terms of vehicle mass, geometry and stiffness. Incompatibility has typically not

been used to describe crash situations in cases where both vehicles and occupants are

identical. In Paper III, scenarios with large differences in terms of passenger compartment

intrusion in one car compared to the other were found around 15° oblique impact angle

combined with small overlap of car front ends. These results needed to be validated and it

was therefore decided that model validation by comparing to physical crash tests was

required.

4.4.2 Method

Two full-scale crash tests were performed in order to validate the crash simulation model,

a 15° oblique angle car-to-car setup, and a rigid barrier test overlapping 25% of the car.

During the analysis of these two crash tests in comparison to simulation results, three major

areas were identified where modelling improvements were required for increased model

validity: front subframe bushings, rim failure and tyre separation. A parameter study was then

performed in order to describe oblique small overlap car-to-car crashes.

4.4.3 Results

Five separate crash categories were distinguished in the simulations. Category A was defined

by at least one of the frontal structures being deformed and thus contributing to energy

absorption in the crash. Category B refers to the set of crashes where the wheels overlap each

other, creating a locking phenomenon which leads to substantial deformations of one of the

identical car models. Categories C and D included situations where the wheel of one car

becomes detached and is pushed along the sill of the same car, leading to un-robust response

in terms of intrusion. Category E comprised sideswipe situations, with a minimal influence

on the struck vehicle front structure and A-pillar. The combination of large intrusions and

large lateral velocity change was only found for the Category B crashes.

4.4.4 Discussion

Predicting rim failure proved to be difficult since a substantial degree of variation is involved

in the physical tests and the borderlines between crash categories do therefore not represent

exact limits. If detachment of the front wheels cannot be achieved in a predictable way,

intrusion may vary substantially. Previous studies have identified that severe injuries often

occur in situations without significant intrusion, suggesting that further development of

restraint systems in combination with improving the deformation modes of frontal structures

may be needed.

24

4.5 Summary of Paper V

4.5.1 Introduction

Studies in Europe as well as the United States into the real-world performance of vehicles

have shown that many belted occupants still sustain severe injuries in frontal crashes without

significant vehicle passenger compartment intrusion. Besides improving the functionality and

robustness of restraint systems, injuries related to the crash pulse can potentially be addressed

by the vehicle front structural response. It has been suggested that adaptive structures can be

used in order to affect the deceleration response in frontal crashes. The front engine subframe

has been identified as a major load path in frontal crashes, important both for self-protection

and crash compatibility. Equipping vehicles with an adaptive detachable front subframe has

the potential to benefit the overall real-world crash performance. The aim of this study was

therefore to quantify the effect an adaptive detachable front subframe has on occupant

loading in car-to-car frontal crashes in a range of lateral offset distances and closing velocity

levels.

4.5.2 Method

A full vehicle model was simplified by removing the majority of the structural components

from a plane rear of the A-pillars in order to perform a large number of simulations. It was

shown that the deceleration response of the simplified model was similar to the original

simulation model at both full overlap and approximately 50% overlap with identical vehicles.

The simplified crash model was also shown to represent the crash pulse shape of vehicles

with similar stopping distance in physical crash tests. A simulation matrix was established to

study the effect of the vehicle deceleration pulse on two simplified crash severity indicators

called Volvo Pulse Index (VPI) and Occupant Load Criterion (OLC).

4.5.3 Results

A high level of correlation was found between longitudinal velocity change and both crash

severity indicators, VPI and OLC, for lateral offset up to and including 1,000 mm. Using this

relationship, an equivalent velocity change reduction (EVCR) was calculated, indicating the

required velocity change reduction in order to achieve the same reduction in crash severity

indicators in the base, passive model, as was achieved by actively detaching the subframe.

The greatest reduction in crash severity gained by releasing the subframe was predicted to be

equivalent to a 44% ΔVx reduction for the VPI model, and 31% for the OLC model. As an

average of the considered crash scenarios, the results based on the VPI model suggest 28%

relative EVCR compared to 18% for the OLC model.

4.5.4 Discussion

The simplification of the vehicle model resulted in a small influence on the crash pulse shape

and it was therefore considered an adequate substitute for the original model. If the subframe

cannot be detached by both vehicles as assumed in this study, the effect on the stopping

distance will be reduced for both vehicles. The VPI and OLC models represent fundamentally

different assumptions on the restraint system characteristics. OLC assumes a perfectly

adaptive restraint system whereas VPI assumes a linearly increasing deceleration without any

limitation on relative displacement. A typical restraint system may therefore be considered as

a combination of the two models and the real-world effect of an adaptive subframe is most

likely within the range given by the VPI and OLC models. Using the front subframe for

structural adaptivity was demonstrated to have a considerable effect on the full-vehicle crash

pulse shape. Benefits over previously proposed solutions were seen, since additional

packaging space or modifications to the frontal longitudinal members would not be required

and the concept may be implementable in an already existing vehicle structure.

25

5 General discussion

Although passenger car crashworthiness has improved tremendously over the years, there are

still issues that should be considered in order to preserve the positive development in the

reduction of the number of fatal and severe injuries in passenger car crashes. Improved CAE

modelling techniques bring new possibilities for predicting the real-world crashworthiness

before vehicles are produced. The methodology presented in this thesis is one such approach

towards vehicle designs that can constitute one of the enablers for reaching a common vision

of zero fatalities and severe injuries in road traffic. The methods and solutions in this thesis

are not expected to be directly applicable to all aspects of robust crashworthiness in passenger

cars. Nevertheless, the suggested approaches adopted in the present thesis are intended as

inspiration to utilise the power of state-of-the art simulation technology to chart future safety

strategies.

Model validity is an issue that always needs to be addressed when findings based on FE

simulation results are discussed. During development of the methodology presented in Paper

III, it became obvious that model updates were needed in order to obtain realistic model

response in terms of separation of non-structural parts, such as wheels and wheel suspension.

Judging by crash tests in small overlap situations with rigid barriers or other cars, the

structural response can differ remarkably compared to standardised barrier crash tests with

deformable honeycomb elements or full-width rigid barriers. Therefore a first step towards

better simulation models was taken and presented in Paper IV. Additional modelling

improvements are expected to be required in order to advance with the proposed

methodology. This is in line with the general trend that can be seen in CAE, since shorter

development time and fewer physical tests during the design phase of passenger cars require

greater confidence in simulation models. Improved reliability of CAE models must, however,

be combined with an appropriate set of load cases that guarantees robust structural behaviour

in real-world situations.

26

5.1 Discussion of results with respect to crash opponent

The compatibility domain established in Paper III was used to provide an overview of how

different crash scenarios compare to each other in terms of intrusion or deceleration in two

identical passenger cars, as illustrated in Figure 11. It is suggested that the usage of the

compatibility domain could be extended to assess factors such as vehicle mass, size, stiffness

in relation to crash configuration factors, such as oblique angle and horizontal or vertical

misalignment. In Paper III, a specific area was used to represent the overall passenger

compartment intrusion levels. For other vehicle designs, such a representative intrusion area

may be more difficult to establish, which would require additional intrusion areas to be

monitored in order to compare intrusion response in a set of crash situations. This will

introduce a more complex comparison of crash situations and may lead to conflicting results,

i.e., one situation leading to large intrusions in one measurement area but small intrusions in

another area, and vice versa for a different situation.

Figure 11. Compatibility domain described in Paper III.

The preliminary findings based on the current state of FE models support previous work on

the significance of small overlap situations (Eichberger et al. 2007, Brumbelow and Zuby

2009). In relation to all the crash scenarios evaluated in Paper III, small overlap situations

stand out as extreme in terms of passenger compartment intrusion. Whether the same

tendencies would be found for a different car design is unknown, however, the methodology

should be readily applicable for this type of studies. Furthermore, there are many

combinations of car designs that could be investigated in terms of car-to-car crash response.

The wheels and wheel suspension components have been identified as important load paths

for small overlap crashes. The design and strength of these components are therefore likely to

have a considerable effect on the structural response and consequently the categorisation into

sub-types of small overlap crashes described in Paper IV.

27

5.2 Discussion of results with respect to crash scenario

As previously identified, robust structural behaviour constitutes a foundation for restraint

system functionality. Further, optimising passenger car structures towards a limited set of

crash load cases may introduce sub-optimisation in a wider perspective on crashworthiness.

The concepts behind robust structural response should therefore be applicable to other crash

scenarios, such as side and rear crashes. The compatibility domain suggested in Paper III may

not be directly applicable in front-to-side crashes since injury risk is normally significantly

higher in the struck vehicle than the striking vehicle in a front-to-side crash (Summers et al.

2003). For front-to-rear crashes, passenger compartment intrusion in the striking vehicle

could be plotted versus relevant intrusion measurements in the struck vehicle. In this way, the

compatibility domain may be employed to compare different structural concepts in order to

ultimately balance injury risk in both the striking and the struck vehicle.

For the car-to-car simulations performed in Paper III, a rough probability distribution of

oblique angles from Eichberger et al. (2007) were taken into account when setting up a

simulation matrix. This indicated that most crashes occur at an oblique angle below 10°

although the number of cases considered in that study was limited to twenty cases. Therefore,

in order to obtain a comprehensive view on frontal crashes, it was decided to cover the

complete range of oblique angles between 0 and 45°.

All studies in this thesis were focused on horizontal alignment in frontal car-to-car crashes;

vertical misalignment was not addressed at all. Based on previous work, there appears to be

agreement that vertical misalignment increases the risk of under riding and increased

intrusion (Baker et al. 2008, Mizuno and Arai 2010). This is obviously an important factor to

consider when striving for improved compatibility in car-to-car crashes. In terms of

methodology development however, it was decided not to attempt to incorporate vertical

misalignment as well.

When considering dimensioning crash scenarios for car structural design, road infrastructure

design also plays an important role. An example of this is lane separation actions as pointed

out by Eugensson et al. (2011). This type of road safety countermeasure can be used in order

to avoid head-on collisions on roads with speed limits above a certain level and thereby

reducing the crashworthiness design volume described in Figure 2.

Small overlap scenarios stand out as particularly demanding for car structures in terms of

passenger compartment intrusion. Crash scenarios combining small horizontal structural

overlap and oblique angles between 10° and 20° are suggested as dimensioning for intrusion

in frontal car-to-car crashes. This finding should be confirmed with alternative vehicle

designs using the methodology described in Paper III. Further, structural countermeasures for

small overlap situations were not suggested or within the scope of this thesis. However, to

prevent opposing front wheels to lock, it has been proposed to actively turn the front wheel

toe-in in order to create a sliding plane from which the crash opponent could be diverted

(Winkler et al. 2001). This action may require a substantial wheel rotation angle before

achieving a positive effect. Such action must also be well balanced with the risk of a second

impact in cases where turning of the wheels is successful in avoiding a first impact.

28

5.3 Discussion of results with respect to crash energy

For the studies conducted in this thesis, energy levels were estimated and probability

distributions were not taken into account. For all studied scenarios, exposing the vehicle

structure to increased energy resulted in either higher accelerations, larger intrusions, or both.

From a car designer’s perspective, the crash energy should be regarded as a framework for

which structural integrity must be ensured. Again, road infrastructure can contribute

significantly to Vision Zero, by setting speed limits to appropriate levels in order to keep

crash energy levels within the crashworthiness design volume.

During the process of active safety systems becoming increasingly efficient in terms of

avoiding and mitigating crashes, virtual tools will be essential in order to ensure that

occupants will be protected in any crashes that still occur. These active safety systems will

thus reduce the crashworthiness design volume as illustrated in Figure 12 which means

reducing the energy level that passive safety systems need to be designed for. In a future

without crashes, passive safety systems would consequently become obsolete.

Figure 12. Visualisation of crashworthiness design volume, and how the number of design situations for

passive safety can be reduced by active safety systems.

When investigating an adaptive front subframe in Paper V, it was found that such a solution

could not be recommended at a closing velocity above 100 km/h for identical vehicles. This

recommendation was based on intrusions being increased above the level of the base model

with a non-releasing subframe. Furthermore, this result indicates that an adaptively detaching

subframe may not be able to reduce fatalities and severe injuries at closing velocities above

100 km/h. An alternative approach to address these high-speed large overlap situations may

be to increase the frontal structural force, e.g., by pressurised structures as proposed by

Pipkorn and Håland (2005). In this way, it would be possible to change a stiff response in the

end of the crash from the integrity of the passenger compartment to a more even deceleration

over the same distance. However, there is undoubtedly an upper limit of velocity change that

can be tolerated by the average car occupant without sustaining severe or fatal injuries. It was

concluded in another study by Pipkorn et al. (2005) that occupant protection in frontal

crashes up to 80 km/h is feasible but may require extended distances available for absorption

of occupant kinetic energy.

The initial studies reported in Paper I and II were directed towards the applicability of

adaptive front structures for reducing the risk of AIS1 neck injuries in frontal crashes.

Although indications of harmful crash pulse characteristics were found in literature (Kullgren

29

et al. 2000, Jakobsson 2004), no clear guidelines towards safer crash pulses with regards to

AIS1 neck injuries were found. Therefore, no further studies were conducted within the scope

of this thesis in order to use adaptive structures for this specific purpose. It is expected,

however, that when crash pulse shape recommendations for reduced AIS1 neck injury risk

can be given, adaptive structures may be one way to realise such crash pulses.

30

6 Conclusions

The following conclusions were drawn:

I. Structural adaptivity can reduce deceleration levels significantly in frontal crashes.

This, however, requires that the available front structure crush length is not

exceeded.

II. In order to achieve efficient structural adaptivity, global load paths need to be

modified. When the material strength of the most significant energy-absorbing

frontal structures were changed by ±50%, only moderate changes to the passenger

compartment deceleration were observed.

III. A methodology for assessing structural robustness based on FE simulation models

was proposed. This methodology points out large overlap situations as extreme in

terms of occupant deceleration loading, whereas oblique small overlap situations

were predicted to be extreme in terms of passenger compartment intrusion.

IV. The results obtained with an updated model based on validation against physical

crash tests exhibited large variations in intrusion response as a function of the

input variables lateral offset and oblique angle. Car-to-car front crash situations

where the front wheels lock up were suggested as critical for occupant safety since

this type of situation combined large intrusion with high crash longitudinal crash

pulse severity and large lateral velocity change.

V. In the last study, the methodology in Paper III was used as a starting point for

addressing the possibility of structural adaptivity to reduce crash pulse severity. It

was shown that a detachable front subframe could be used at closing velocities

below 100 km/h. Detaching the front subframe was suggested to reduce the crash

severity equivalent to a velocity change reduction of up to 44%.

6.1 Lessons learned

Some experiences related to the number of crash simulations performed with the standard

models in Paper III as opposed to the updated, more detailed models in Paper IV were gained.

When searching for dimensioning scenarios, it would have been beneficial to use a fixed set

of initial velocities to save simulation effort in the first stage. The car-to-car model validation

work conducted in Paper IV provided several important insights into how details of the

models (Centeno G. 2009, Dharwadkar 2011) affect the structural response. Therefore, a

larger number of simulations could have been based on the updated Paper IV models to scan

the crash configuration domain, also denoted crash scenario matrix.

The detail level of the FE model used for Paper II (16,000 elements) appears to be a limiting

factor when compared to the model used for Paper III and IV (>2,000,000 elements). Element

size has a direct effect on a model’s ability to capture buckling modes of structural

components; a coarse model can simply not capture some deformation modes that a finer

model can. Nevertheless, usage of simplified or coarse models should not be disregarded

since using these can be an effective way of sorting between high-level decisions. For

instance, the studies on adaptive front longitudinal rails, or side members, in Paper II

indicated that significant load paths would need to be affected in order to considerably affect

the vehicle crash pulse. This was later supported when the concept of an adaptive subframe,

which is a major load path in frontal crashes, was demonstrated in Paper V.

31

Important strategic decisions should be made regarding on how to spend simulation

resources. On the one hand, crash models can be made increasingly detailed in order to

predict crash performance with higher precision in a constant number of load cases,

corresponding to moving right on the horizontal axis in Figure 13. On the other hand, as was

demonstrated in Paper III, the number of load cases can be increased with a constant or even

reduced model detail level, corresponding to moving up on the vertical axis in Figure 13.

Based on studies conducted for this thesis, it is not obvious which choice is more effective in

terms of designing for real-world crashworthiness. A trade-off is needed between increasing

the model detail level and the number of situations to evaluate. When a new computer system

generation becomes available, as illustrated in Figure 13, which enables a larger number of

computations to be performed, this trade-off needs to be considered. It is expected that this

balance will be affected by which part of the development process that is considered, i.e., at

earlier phases of development a lower model detail level may be preferable while exploring a

larger number of situations.

Figure 13. Schematic view on trade-off between model detail level and number of situations to evaluate

during design process of passenger cars. Alternative development paths marked by dotted

arrows.

32

6.2 Suggestions for future studies and development work

Based on the research performed for this thesis, there appears to be three obvious paths for

continued work. Table 1 was therefore updated with an outline of these paths in Table 3.

The first path would be useful for going into further details of passive safety issues and

vehicle structures in frontal crashes as illustrated in Table 3. Issues such as vertical

misalignment and robustness in car-to-car crashes with dissimilar vehicles that were not

covered by the present studies could be addressed. Furthermore, connecting crash simulation

setups more closely to accident data including the probability of crash scenarios occurring

may be one way of advancing towards robust structural design. A further application could be

single vehicle frontal crashes with fixed objects or large animals.

A second path deals with the generalisation of the methodology to encompass other crash

scenarios such as rear and side crashes. Using a set of bullet vehicle models, a structural

robustness map could be established in a similar fashion as was made for frontal crashes.

A third, and perhaps the most innovative, way forward is looking into integrated safety for

frontal crashes, i.e., the interaction between active and passive safety systems aiming to

reduce accidents and injuries. The first step in this work is to combine the methodology for

structural robustness described in this thesis with detailed occupant models. This could

possibly quantify injury risks related to intrusions and decelerations on a considerably more

detailed level than can be achieved using intrusion measurements and crash severity

indicators such as VPI (ISO 2012) or OLC (Stein et al. 2011).

Table 3. Overview of how methods and solutions could be expanded to include additional crash

scenarios and safety systems. Numbers refer to alternative future work paths.

As continued focus is being placed on energy efficiency, reducing vehicle mass is one of the

options for producing cars that use less fuel. The safety consequences of a future lighter

vehicle fleet in the United States has been thoroughly analysed by NHTSA (2013). If lighter

materials such as ultra-high strength steel, aluminium or composites are implemented, this

must be done with robust structural response in mind. This kind of assessments during the

development phase requires both detailed models showing the structural response and broad

assessments of crash situations.

33

Another potential opportunity to increase energy efficiency fuel is electric vehicles, an area

where several research projects have been initiated, e.g., ELVA (2013) and EVERSAFE

(2013). The transition to alternative powertrains may introduce challenges, as well as

opportunities in terms of the structural response when future vehicles are involved in crashes.

Furthermore, safety design experience built around conventional vehicles may not be directly

applicable to vehicles incorporating alternative powertrains. Therefore, it is proposed that it

would be beneficial to apply the methodology presented in Paper III when designing electric

vehicles offering robust structural response.

34

7 References

Arai, Y., Yamazaki, K., Mizuno, K. and Kubota, H. (2007). Full-Width Tests To Evaluate

Structural Interaction. International Technical Conference on the Enhanced Safety of

Vehicles, Lyon, France.

Austin, R. (2005). Vehicle Aggressiveness in Real World Crashes. International Technical

Conference on the Enhanced Safety of Vehicles, Washington D.C., U.S.A.

Baker, B. C., Nolan, J. M., O'Neill, B. and Genetos, A. P. (2008). Crash compatibility

between cars and light trucks: Benefits of lowering front-end energy-absorbing structure in

SUVs and pickups. Accident Analysis & Prevention, Vol. 40, No. 1, pp. 116-125.

Barbat, S. (2005). Status of Enhanced Front-To-Front Vehicle Compatibility Technical

Working Group Research and Commitments. International Technical Conference on the

Enhanced Safety of Vehicles, Washington D.C., U.S.A.

Brewer, J., Patel, S., Summers, S., Prasad, A. and Mohan, P. (2011). Accuracy of AHOF400

with a Moment-Measuring Load Cell Barrier. International Technical Conference on the

Enhanced Safety of Vehicles, Washington, D.C., U.S.A.

Brumbelow, M. and Zuby, D. (2009). Impact and injury patterns in frontal crashes of

vehicles with good ratings for frontal crash protection. International Technical Conference

on the Enhanced Safety of Vehicles, Stuttgart, Germany.

Buzeman-Jewkes, D. G., Viano, D. C. and Lövsund, P. (1999). A Multi-body Integrated

Vehicle-Occupant Model for Compatibility Studies in Frontal Crashes. Journal of Crash

Prevention and Injury Control, Vol. 1, No. 2, pp. 143 - 154.

Buzeman-Jewkes, D. G., Viano, D. C. and Lövsund, P. (2000). Occupant Risk, Partner Risk

and Fatality Rate in Frontal Crashes: Estimated Effects of Changing Vehicle Fleet Mass in

15 Years. Journal of Crash Prevention and Injury Control, Vol. 2, No. 1, pp. 1 - 10.

Centeno G., O. J. (2009). Finite Element Modeling of Rubber Bushing for Crash Simulation -

Experimental Tests and Validation. MSc thesis, Lund University, Lund, Sweden.

Chauvel, C., Faverjon, G., Bertholon, N., Cuny, S. and Delannoy, P. (2011). Self-Protection

and Partner-Protection for New Vehicles (UNECE R94 Amendment). International

Technical Conference on the Enhanced Safety of Vehicles, Washington, D.C., U.S.A.

Conroy, C., Tominaga, G. T., Erwin, S., Pacyna, S., Velky, T., Kennedy, F., Sise, M. and

Coimbra, R. (2008). The influence of vehicle damage on injury severity of drivers in head-

on motor vehicle crashes. Accident Analysis & Prevention, Vol. 40, No. 4, pp. 1589–1594.

Craig, K. J., Stander, N., Dooge, D. A. and Varadappa, S. (2005). Automotive

crashworthiness design using response surface-based variable screening and optimization.

Engineering Computations, Vol. 22, No. 1-2, pp. 38-61.

DaCoTA (2011). Annual Statistical Report 2010. Retrieved June 27, 2013, from

http://www.dacota-project.eu/deliverables.html.

35

Delannoy, P. and Diboine, A. (2001). Structural Front Unit Global Approach. International

Technical Conference on the Enhanced Safety of Vehicles, Amsterdam, The Netherlands.

Delannoy, P., Martin, T. and Castaing, P. (2005). Comparative Evaluation of Frontal Offset

Tests to Control Self and Partner Protection. International Technical Conference on the

Enhanced Safety of Vehicles, Washington D.C., U.S.A.

Delannoy, P., Martin, T., Meyerson, S., Summers, L. and Wiacek, C. (2007). PDB Barrier

Face Evaluation by DSCR and NHTSA’s Joint Research Program. International Technical

Conference on the Enhanced Safety of Vehicles, Lyon, France.

Dharwadkar, N. A., Krishna Prashant (2011). Modelling of Engine Suspension Components

for Crash Simulations. MSc thesis, Chalmers University of Technology, Gothenburg,

Sweden.

Edwards, M. (2009). Development of a High Deceleration Full Width Frontal Impact Test for

Europe. International Technical Conference on the Enhanced Safety of Vehicles, Stuttgart,

Germany.

Edwards, M., Cuerden, R. and Davies, H. (2007). Current Status of the Full Width

Deformable Barrier Test. International Technical Conference on the Enhanced Safety of

Vehicles, Lyon, France.

Edwards, M., Davies, H. and Hobbs, A. (2003a). Development of Test Procedures and

Performance Criteria to Improve Compatibility in Car Frontal Collisions. International

Technical Conference on the Enhanced Safety of Vehicles, Nagoya, Japan.

Edwards, M., Davies, H., Thompson, A. and Hobbs, A. (2003b). Development of test

procedures and performance criteria to improve compatibility in car frontal collisions.

Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile

Engineering, Vol. 217, No. 4, pp. 233-245.

Edwards, M., Happian-Smith, J., Davies, H., Byard, N. and Hobbs, A. (2001). The Essential

Requirements for Compatible Cars in Frontal Collisions. International Technical

Conference on the Enhanced Safety of Vehicles, Amsterdam, the Netherlands.

Eichberger, A., Schimpl, W., Werber, C. and Steffan, H. (2007). A new crash test

configuration for car-to-car frontal collisions with small lateral overlap. International

Journal of Crashworthiness, Vol. 12, No. 2, pp. 93 - 100.

Elmarakbi, A. M. and Zu, J. W. (2006). Crash analysis and modeling of two vehicles in

frontal collisions using two types of smart front-end structures: an analytical approach

using IHBM. International Journal of Crashworthiness, Vol. 11, No. 5, pp. 467 - 483.

ELVA (2013). Advanced Electric Vehicle Architectures. Retrieved June 5, 2013, from

http://www.elva-project.eu/.

Eugensson, A., Ivarsson, J., Lie, A. and Tingvall, C. (2011). Cars are Driven on Roads, Joint

Visions and Modern Technologies Stress the Need for Co-Operation. International

Technical Conference on the Enhanced Safety of Vehicles, Washington, D.C., U.S.A.

36

European Commission (2013). EU road fatalities. Retrieved March 4, 2013, from

http://ec.europa.eu/transport/road_safety/pdf/observatory/trends_figures.pdf.

EVERSAFE (2013). Everyday Safety for Electric Vehicles. Retrieved June 5, 2013, from

http://www.eversafe-project.eu/.

Farmer, C. M. (2005). Relationships of frontal offset crash test results to real-world driver

fatality rates. Traffic Injury Prevention, Vol. 6, No. 1, pp. 31 - 37.

FIMCAR (2011a). D1.1 Report detailing the analysis of national accident databases.

Retrieved March 5, 2013, from www.fimcar.eu/results/.

FIMCAR (2011b). D1.2 Benefit analysis. Retrieved May 24, 2013, from

http://www.fimcar.eu/results/.

FIMCAR (2011c). D2.2 Updated Off-set Test Protocol. Retrieved May 28, 2013, from

http://www.fimcar.eu/results/.

FIMCAR (2011d). D4.1 MPDB Test Protocol. Retrieved March 22, 2013, from

http://www.fimcar.eu/results/.

FIMCAR (2011e). D6.3 Documentation for a frontal impact and compatibility assessment

approach: Part I. Retrieved March 22, 2013, from http://www.fimcar.eu/results/.

Fredette, M., Mambu, L. S., Chouinard, A. and Bellavance, F. (2008). Safety impacts due to

the incompatibility of SUVs, minivans, and pickup trucks in two-vehicle collisions. Accident

Analysis & Prevention, Vol. 40, No. 6, pp. 1987-1995.

Fujii, S., Fukushima, M., Abe, A., Ogawa, S., Fujita, H., Sunakawa, T. and Tanaka, Y.

(2003). Vehicle Front Structure in Consideration of Compatibility. International Technical

Conference on the Enhanced Safety of Vehicles, Nagoya, Japan.

Gabler, H. C. and Hollowell, W. T. (2000). The Crash Compatibility of Cars and Light

Trucks. Journal of Crash Prevention and Injury Control, Vol. 2, No. 1, pp. 19 - 31.

Greenwell, N. K. (2012). Evaluation of the Enhancing Vehicle-to-Vehicle Crash

Compatibility Agreement: Effectiveness of the Primary and Secondary Energy-Absorbing

Structures on Pickup Trucks and SUVs. U.S. Department of Transportation, Report no. HS

811 621.

Hirayama, S., Watanabe, T., Obayashi, K. and Okabe, T. (2007). Second Report of Research

on Stiffness Matching Between Vehicles for Frontal Impact Compatibility. International

Technical Conference on the Enhanced Safety of Vehicles, Lyon, France.

Huang, H., Siddiqui, C. and Abdel-Aty, M. (2011). Indexing crash worthiness and crash

aggressivity by vehicle type. Accident Analysis & Prevention, Vol. 43, No. 4, pp. 1364-

1370.

ISO (2012). ISO/PDTR 12353-3: Road vehicles - Traffic accident analysis - Part 3:

Guidelines for the interpretation of recorded crash pulse data to determine impact severity,

International Organization for Standardization.

37

Jakobsson, L. (2004). Evaluation of impact severity measures for AIS 1 neck injuries in

frontal impacts using crash recorder data. International Journal of Crashworthiness, Vol. 9,

No. 1, pp. 105-111.

Jakobsson, L., Kling, A., Lindman, M., Wågström, L., Axelson, A. and Broberg, T. (2013a).

Severe Partial Overlap Crashes – A Methodology Representative of Car-to-Car Real World

Frontal Crash Situations. International Technical Conference on the Enhanced Safety of

Vehicles, Seoul, South Korea.

Jakobsson, L., McInally, G., Axelson, A., Lindman, M., Kling, A., Broberg, T., Fermér, M.

and Wågström, L. (2013b). Severe Frontal Collisions with Partial Overlap – Two Decades

of Car Safety Development. SAE World Congress, Detroit, Michigan, U.S.A.

Jenefeldt, F. and Thomson, R. (2004). A methodology to assess frontal stiffness to improve

crash compatibility. International Journal of Crashworthiness, Vol. 9, No. 5, pp. 475 - 482.

Johansson, R. (2009). Vision Zero–Implementing a policy for traffic safety. Safety Science,

Vol. 47, No. 6, pp. 826-831.

Khattab, A. A. E. (2011). Investigation of an Adaptable Crash Energy Management System to

Enhance Vehicle Crashworthiness. PhD thesis, Concordia University, Montreal, Quebec,

Canada.

Kullgren, A., Krafft, M., Nygren, Å. and Tingvall, C. (2000). Neck injuries in frontal

impacts: influence of crash pulse characteristics on injury risk. Accident Analysis &

Prevention, Vol. 32, No. 2, pp. 197-205.

Kullgren, A., Lie, A. and Tingvall, C. (2001). Mass Data Evaluation of the Importance of

Structural and Mass Related Aggressivity. International Technical Conference on the

Enhanced Safety of Vehicles, Amsterdam, the Netherlands.

Kullgren, A., Lie, A. and Tingvall, C. (2010). Comparison between Euro NCAP test results

and real-world crash data. Traffic Injury Prevention, Vol. 11, No. 6, pp. 587 - 593.

Kullgren, A. and Ydenius, A. (1998). Frontal Impacts with Small Partial Overlap: Real Life

Data from Crash Recorders. International Technical Conference on the Enhanced Safety of

Vehicles, Windsor, Ontario, Canada.

Kühn, D. M., Hummel, T. and Bende, J. (2013). Small-Overlap Frontal Impacts involving

Passenger Cars in Germany. International Technical Conference on the Enhanced Safety of

Vehicles, Seoul, South Korea.

Lee, K.-H. and Park, G.-J. (2001). Robust optimization considering tolerances of design

variables. Computers & Structures, Vol. 79, No. 1, pp. 77-86.

Lindman, M. (2012). Volvo Cars fatal crashes. Internal report.

Lindquist, M., Hall, A. and Björnstig, U. (2004). Car structural characteristics of fatal

frontal crashes in Sweden. International Journal of Crashworthiness, Vol. 9, No. 6, pp. 587

- 597.

38

Lund, A. K. and Nolan, J. M. (2003). Changes in Vehicle Designs from Frontal Offset and

Side Impact Crash Testing. SAE Paper 2003-01-0902.

Lönn, D., Bergman, G., Nilsson, L. and Simonsson, K. (2011). Experimental and finite

element robustness studies of a bumper system subjected to an offset impact loading.

International Journal of Crashworthiness, Vol. 16, No. 2, pp. 155-168.

Lönn, D., Fyllingen, Ø. and Nilssona, L. (2010). An approach to robust optimization of

impact problems using random samples and meta-modelling. International Journal of

Impact Engineering, Vol. 37, No. 6, pp. 723-734.

Lönn, D., Öman, M., Nilsson, L. and Simonsson, K. (2009). Finite element based robustness

study of a truck cab subjected to impact loading. International Journal of Crashworthiness,

Vol. 14, No. 2, pp. 111-124.

Marklund, P. O. and Nilsson, L. (2001). Optimization of a car body component subjected to

side impact. Structural and Multidisciplinary Optimization, Vol. 21, No. 5, pp. 383-392.

Meyerson, S., Wiacek, C., Delannoy, P. and Robert, G. (2009). Evaluation of Advanced

Compatibility Frontal Structures Using the Progressive Deformable Barrier (PDB).

International Technical Conference on the Enhanced Safety of Vehicles, Stuttgart,

Germany.

Mizuno, K. and Arai, Y. (2010). The effectiveness of matching front rail heights in car-to-car

crashes. International Journal of Crashworthiness, Vol. 15, No. 6, pp. 571 - 579.

Mizuno, K., Arai, Y., Yamazaki, K., Yonezawa, H. and Notsu, M. (2005). Test procedures

for vehicle compatibility evaluation. International Journal of Crashworthiness, Vol. 10, No.

5, pp. 473 - 481.

Mizuno, K. and Kajzer, J. (1999). Compatibility problems in frontal, side, single car

collisions and car-to-pedestrian accidents in Japan. Accident Analysis & Prevention, Vol.

31, No. 4, pp. 381-391.

Motozawa, Y. and Kamei, T. (2000). A New Concept for Occupant Deceleration Control in a

Crash. SAE Paper 2000-01-0881.

Motozawa, Y., Tsuruta, M., Kawamura, Y. and Noguchi, J. (2003). A New Concept for

Occupant Deceleration Control in a Crash - Part 2. SAE Paper 2003-01-1228.

Mueller, B. C., Sherwood, C., Nolan, J. M. and Zuby, D. S. (2013). Repeatability of IIHS

Frontal Small Overlap Frontal CrashTests. International Technical Conference on the

Enhanced Safety of Vehicles, Seoul, South Korea.

NHTSA (2009). Fatalities in frontal crashes despite seat belts and air bags - review of all

CDS cases model and calendar years 2000-2007: 122 fatalities. U.S. Department of

Transportation, Report no. HS 811 202.

NHTSA (2013). Corporate Average Fuel Economy. Retrieved June 5, 2013, from

http://www.nhtsa.gov/fuel-economy.

39

Nolan, J. M. and Lund, A. K. (2001). Frontal Offset Deformable Barrier Crash Testing and

its Effect on Vehicle Stiffness. International Technical Conference on the Enhanced Safety

of Vehicles, Amsterdam, The Netherlands.

Nusholtz, G., Xu, L., Velez, E. L., Hsu, T. and Kochekseraii, S. B. (2009). Reproducibility of

AHOF400 and KW400. International Technical Conference on the Enhanced Safety of

Vehicles, Stuttgart, Germany.

Ostrowski, M., Holnicki-Szulc, J. and Griskevicius, P. (2007). Adaptive Crashworthiness of

Front-End Structure of Motor Vehicles. 2007 SAE World Congress, Detroit, Michigan,

SAE International.

Patel, S., Smith, D., Prasad, A. and Mohan, P. (2007). NHTSA’s recent vehicle crash test

program on compatibility in front to front impacts. International Technical Conference on

the Enhanced Safety of Vehicles, Lyon, France.

Pipkorn, B., Fredriksson, R. and Olsson, J. (2007). Bumper Bag for SUV to Passenger

Vehicle Compatibility and Pedestrian Protection. International Technical Conference on the

Enhanced Safety of Vehicles, Lyon, France.

Pipkorn, B. and Håland, Y. (2005). Proposed variable stiffness of vehicle longitudinal frontal

members. International Journal of Crashworthiness, Vol. 10, No. 6, pp. 603 - 608.

Pipkorn, B. and Kullgren, A. (2009). Effects of active structures on injuries in medium

severity frontal impacts. International Technical Conference on the Enhanced Safety of

Vehicles, Stuttgart, Germany.

Pipkorn, B., Lundström, J. and Ericsson, M. (2011). Safety and Vision Improvements by

Expandable A-Pillars. International Technical Conference on the Enhanced Safety of

Vehicles, Washington, D.C., U.S.A.

Pipkorn, B., Mellander, H. and Håland, Y. (2005). Car Driver Protection at Frontal Impacts

up to 80 km/h. International Technical Conference on the Enhanced Safety of Vehicles,

Washington D.C., U.S.A.

Planath, I., Norin, H. and Nilsson, S. (1993). Severe Frontal Collisions with Partial Overlap -

Significance, Test Methods and Car Design. SAE Paper 930636.

Rudd, R. W., Bean, J., Cuentas, C., Kahane, C. J., Mynatt, M. and Wiacek, C. (2009). A study

of the factors affecting fatalities of air bag and beltrestrained occupants in frontal crashes.

International Technical Conference on the Enhanced Safety of Vehicles, Washington, D.C.,

U.S.A.

Rudd, R. W., Scarboro, M. and Saunders, J. (2011). A study of the factors affecting fatalities

of air bag and beltrestrained occupants in frontal crashes. International Technical

Conference on the Enhanced Safety of Vehicles, Washington, D.C., U.S.A.

Saito, M., Gomi, T., Taguchi, Y., Yoshimoto, T. and Sugimoto, T. (2003). Innovative Body

Structure for the Self-Protection of a Small Car in a Frontal Vehicle-to-Vehicle Crash.

International Technical Conference on the Enhanced Safety of Vehicles, Nagoya, Japan.

40

Samaha, R. R., Digges, K., Fesich, T. and Authaler, M. (2010). Frontal Crash Testing and

Vehicle Safety Designs: A Historical Perspective Based on Crash Test Studies. SAE Paper

2010-01-1024.

Schram, R. and Versmissen, T. (2007). The Development of a Mobile Deformable Barrier

Test Procedure. International Technical Conference on the Enhanced Safety of Vehicles,

Lyon, France.

Scullion, P., Morgan, R. M., Mohan, P., Kan, C.-D., Shanks, K., Jin, W. and Tangirala, R.

(2010). A Reexamination of the Small Overlap Frontal Crash. Annals of Advances in

Automotive Medicine, Vol. 54, No. 137-148.

Sherwood, C. (2009). Characteristics of Small Overlap Crashes. International Technical

Conference on the Enhanced Safety of Vehicles, Stuttgart, Germany.

Sherwood, C. P., Mueller, B. C., Nolan, J. M., Zuby, D. S. and Lund, A. K. (2013).

Development of a Frontal Small Overlap Crashworthiness Evaluation Test. Traffic Injury

Prevention, Vol. 14, No. sup1, pp. S128-S135.

Stein, M., Friedemann, D., Eisenach, A., Zimmer, H. and Johannsen, H. (2011). Parametric

Modelling of Simplified Car Models for Assessment of Frontal Impact Compatibility. 8th

European LS-DYNA Users Conference, Strasbourg , France.

Subramaniam, K., Verma, M., Nagappala, R., Tedesco, R. and Carlin, L. (2007). Evaluation

of Stiffness Matching Concepts for Vehicle Safety Improvement. International Technical

Conference on the Enhanced Safety of Vehicles, Lyon, France.

Summers, S. M., Hollowell, W. T. and Prasad, A. (2003). NHTSA’s research program for

vehicle compatibility. International Technical Conference on the Enhanced Safety of

Vehicles, Nagoya, Japan.

Teoh, E. R. and Nolan, J. M. (2012). Is Passenger Vehicle Incompatibility Still a Problem?

Traffic Injury Prevention, Vol. 13, No. 6, pp. 585-591.

The Insurance Institute for Highway Safety (2005). Special issue: vehicle incompatibility in

crashes. Status Report Vol. 40, No. 5.

The Insurance Institute for Highway Safety (2009). Special Issue: Car size, weight and

safety. Status Report Vol. 44, No. 4.

Thompson, A., Edwards, M., Goodacre, O., Coley, G., Keigan, M., Broughton, J. and

Cuerden, R. (2007). Adaptive Vehicle Structures for Secondary Safety. Retrieved June 5,

2013, from http://www.transportresearchfoundation.co.uk/Publications/.

Thomson, R., Krusper, A., Avramov, N. and Rachid, K. (2008). The Role Of Vehicle Design

On Structural Interaction. International Crashworthiness Conference, Kyoto, Japan.

Tingvall, C. and Haworth, N. (1999). Vision Zero: an ethical approach to safety and mobility.

6th ITE International Conference Road Safety & Traffic Enforcement: Beyond 2000,

Melbourne, Australia.

41

Uwai, H., Toyosaki, S., Sagawa, K. and Takahashi, N. (2007). A Study on AHOF400 which is

Possible Car to SUV Compatibility Evaluation Method. International Technical Conference

on the Enhanced Safety of Vehicles, Lyon, France.

van der Zweep, C. D., Kellendonk, G. and Lemmen, P. (2005). Evaluation of fleet systems

model for vehicle compatibility. International Journal of Crashworthiness, Vol. 10, No. 5,

pp. 483 - 494.

Watanabe, T., Hirayama, S., Obayashi, K. and Okabe, T. (2005). Research on Stiffness

Matching Between Vehicles for Frontal Impact Compatibility. International Technical

Conference on the Enhanced Safety of Vehicles, Washington D.C., U.S.A.

VC-Compat (2007). VC-Compat Final Technical Report. Retrieved May 28, 2013, from

http://vc-compat.rtdproject.net/.

Verma, M. K., Nagappala, R., Murugan, M. and Tung, Y. J. (2004). Evaluation of structural

parameters for vehicle crash compatibility. International Journal of Crashworthiness, Vol.

9, No. 6, pp. 577 - 586.

Versmissen, T., Schram, R. and McEvoy, S. (2007). The development of a load sensing

trolley for frontal offset testing. International Journal of Crashworthiness, Vol. 12, No. 3,

pp. 235 - 245.

Winkler, S., Bernd, F., Wilhelm, B. and Hermann, S. (2001). Sliding Collisions in Case of

Frontal Crash with Small Lateral Offset. Graz, Austria. Retrieved March 4, 2013 from

http://www.tuevsued.de/uploads/images/1134986822332304384899/Winkler.pdf.

Witteman, W. J. (2005). Adaptive frontal structure design to achieve optimal deceleration

pulses. International Technical Conference on the Enhanced Safety of Vehicles,

Washington D.C., U.S.A.

Witteman, W. J. and Kriens, R. (2001). The necessity of an adaptive vehicle structure to

optimize deceleration pulses for different crash velocities. International Technical

Conference on the Enhanced Safety of Vehicles, Amsterdam, The Netherlands.

Volvo Car Corporation, Saab Automobile, Autoliv Development AB, Volvo 3P, Scania CV

AB and Chalmers tekniska högskola AB (2010). Kompatibilitet mellan fordon, Program för

fordonsforskning, Projekt 2005-00776.

World Health Organization (2013). Global Status Report on Road Safety: Time for Action.

Retrieved March 4, 2013, from

http://whqlibdoc.who.int/publications/2009/9789241563840_eng.pdf.

Yonezawa, H., Hosokawa, N., Mizuno, K. and Kiuchi, S. (2008). Evaluation of crash

compatibility based on car-to-car crash tests. International Journal of Vehicle Safety, Vol.

3, No. 1, pp. 81-90.